Finding Lane Lines on the Road

---

Finding Lane Lines on the Road

The goals / steps of this project are the following:

• Make a pipeline that finds lane lines on the road
• Reflect on your work in a written report

---

1. Description of the pipeline

The top level function of my pipeline is lane_detector_from_image(). It consists of x steps

• Convert the RGB image to gray scale using the V channel of HSV space
• Find edges in the image. This is acheived by first apply a Gaussian filter with kernel size of 3, followed by performing canny edge detection, and finally mask the edges founds by region_of_interest fileter
• Combine the edges into line segments by using cv2.HoughLinesP function.
• Merge the line segments. See function merge_lines() for detail. In the function, I iterate though the line segments found in the previous step, and group them if they are aligned within a threshold.
• Select the lane lines from the merged line segments. See function select_lane_from_lines for details. The selection is based on two simple huristics:
  • The lane lines vanishes at the center of the image.
  • Left/right lane line has large positive/negative slope.
• Finally, the lane lines are extended to proper length and annotated on the image.

2. Shortcomings with the current pipeline and possible improvements

• When run on a video input, the annotated lane lines are sometimes shakey from frame to frame. This is due to the current pipeline treat each frame in a video as independent input. As an impromvent, we can store the position of lane lines in previous frames and enforce the lane lines to be updated smoothly between frames
• When the lane line is curved, the current detection pipeline will select the portion of lane line with steeper slope. This can be seen in the beginning of solidYellowLeft.mp4. This is not ideal. A better approach would be understand the lane line consists of multiple portions and take the average of them.